LEAD EXPOSURE AND CHILDREN'S INTELLIGENCE
This is the transcript of a keynote address orignally presented at the Second international Occupational Hygiene Association Conference in Hong Kong CO Sponsoered by the American Industrial Hygiene Association
Do low levels of lead in blood cause mental deficit?
Ms Pamela E de Silva MPH* and Dr Allen J Christophers MBBS+
* AMCOSH Occupational Health Services,
5 Macarthur Street, Melbourne, Australia 3002
+ Senior Associate, Department of Pharmacology,
University of Melbourne, Australia
In the current review of the TLV and BEI for lead, the assessment is to be focused upon parental blood lead values and post natal cognitive development, now regarded as the critical adverse effect of lead. It is claimed by some researchers that exposure of the foetus and young children to very low levels of lead causes mental deficit. If this claim is accepted, there will be an extremely difficult occupational hygiene problem around the world as efforts are made to keep the blood lead levels of women lead workers below about 15g/dl. This paper reviews the evidence for this claim.
The first study relating childhood exposure to lead and intelligence in which confounding variables were properly controlled was done by de la Burde and Choate1 in 1972 and many more have been done since then. In all the early studies, lead exposure and intelligence were measured only once and at the same time (cross-sectional studies).
A meta-analysis of these studies has been recently published by Needleman and Gatsonis2 and a highly statistically significant association between lead exposure and mental deficit was found. Few researchers would now dispute this. The review then pointed out that these studies did not establish whether it was the increased lead exposure that caused a lowering of intelligence or a lower intelligence that caused an increase in lead exposure. However, Needleman and Gatsonis then proceeded to argue, on the basis of these studies and of later prospective studies where cohorts of children were followed from birth, that it was the lead exposure that was responsible for the mental deficit and not the mental deficit that was responsible for the lead exposure.
In this paper, we will accept the association between lead exposure and mental deficit as fact; however, we will argue that, since in all of these studies the lead exposure was due to pica, it is not the lead exposure responsible for the mental deficit, but the mental deficit that is responsible for the lead exposure. Placing mental deficit rather than lead exposure as the cause has been referred to by commentators as the theory of reverse causation and, for clarity, we will use this term.
In support of their conclusions, Needleman and Gatsonis discussed the principles to be followed in the drawing of causal inferences under four headings:
(1) time precedence of association,
(2) biological plausibility,
(3) non-spuriousness, and
Of these principles, the latter two relate only to whether or not a causal relation exists, not to its direction. Only the first two principles are concerned with the causal direction. Our argument will therefore be developed under the headings of Time Precedence of Association and Biological Plausibility.
That mental deficit precedes the increased exposure to lead will be argued in two stages. It will be first shown that mental retardation (low intelligence below the normal range) is associated with greatly increased exposure to lead and that the mental retardation usually precedes and therefore is a cause of increased exposure to lead. Then it will be shown that mental deficit (low intelligence within the normal range) is also associated with some increase in lead exposure; but it cannot be established whether the mental deficit
or the lead exposure has precedence in time. However, mental retardation may be regarded as mental deficit taken beyond the normal range and there is no sharp demarcation between the two. It is to be expected that a causal relationship applying to mental retardation will also apply to mental deficit.
Before these arguments are advanced, it is important to explore the relationship of pica to lead exposure and to comment on the assessment of intelligence.
Pica has usually been defined as a habit of ingesting material unsuitable as food when this habit exceeds that which may be expected in normal child development. Here pica will be defined as the habit whether normal or abnormal.
When blood lead levels of children are followed from birth as they were in the Cincinatti study3, it has been found that the level rises rapidly after 6 months of age and reaches a peak at the age of about 21 months and then declines slowly until at about the age of 5 years it has reached a value about two-thirds of its peak value.
Lead exposure in children is determined by three main factors: availability of lead in the environment; mobility of the child; and the habit of pica. Pica is most marked at about 12 months of age and then slowly declines at a time when mobility, which allows children access to environmental lead, is increasing. The peak of lead exposure is therefore at about 21 months of age when the combination of the two factors is greatest.
It has been well known since early this century that pica, defined as abnormal behaviour, is much more common in the mentally retarded; yet, strangely, this fact has not been directly researched and proven until recently.4,5 The relationship of pica to mental ability in children of normal intelligence range has never been directly researched.
Assessment of intelligence
In cases of mental retardation, the diagnosis may be made in early infancy and in the overwhelming majority of cases, the retardation may be ascribed to causes operating either prenatally or neonatally. In cases of mental deficit within the normal range of intelligence, the diagnosis awaits the use of conventional tests of intelligence and the significance of these vary with the age at which they are administered. Bayley Scales of Infant Development have been used as measures of infant intelligence in all of the lead studies involving children up to 3 years old; but Bayley herself found no correlation between intelligence testings, by any of the generally accepted methods, over the first 4 years of life and intelligence at maturity.6 At the age of 5 years correlation rose to 0.06 and to 0.90 at 11-18 years. Other workers have made similar comments.7,8 Thus mental deficit cannot be diagnosed until the age of about 5 years; but the diagnosis gives no hint of when it was caused. The potential for mental deficit may have been present from birth waiting for manifestation for some years.
Time Precedence of Association: Mental retardation as a cause of increased lead exposure
The first study to show that children institutionalised for mental retardation had raised blood lead levels was that of Moncrieff et al (1964)9. They compared the levels in 80 normal children with those of 122 children admitted to hospital for the investigation of mental retardation. Only 2 of the 80 normal children had a blood lead level above 36g/dl; whereas 55 of the 122 retarded children had blood lead levels greater than this, and 7 of these 122 had a blood lead level above 60g/dl. Several years later, Bicknell et al (1968)10 reported another study of 27 mentally retarded children. The mean blood lead level of the group was 48g/dl. The authors made these pertinent comments on the group of 27 children with severe pica:
The cause of the subnormality was not known in 12 cases; perinatal difficulties sufficient to cause brain damage had occurred in 5; a diagnosis had been made in 10 including one case in which lead poisoning was definitely a causal factor. Of the 12 cases in which a specific diagnosis had not been made, brain damage had been diagnosed in 9 before the age of 12 months and prior to the onset of walking or standing which makes lead poisoning an unlikely cause of primary brain damage. In one the delay in walking raised the question of mental retardation which was subsequently confirmed and in two mental retardation became apparent after walking had been achieved.
Bicknell et al established that, in the overwhelming majority of cases, the mental retardation had time precedence over the lead poisoning and therefore should be regarded as its cause. Since lead poisoning and lead exposure occur at almost exactly the same ages, a corollary is that the mental retardation had time precedence over the lead exposure and therefore should be regarded as its cause.
For mentally retarded children, the association of decreased mental ability with increased lead intake cannot be interpreted as the increased lead intake causing the decreased mental ability as the latter has priority in time. The causal chain is irrefutable. The mental retardation is associated with the increase in pica and the increase in pica then causes the increase in lead intake.
Time Precedence of Association: Mental deficit as a cause of increased lead exposure
In the case of mental deficit within the normal range of intelligence, the accepted facts are very similar to those applying to mental retardation. That there is an association between mental deficit and lead exposure should be accepted and is the corner stone of the argument made by Needleman and Gatsonis as outlined above. The association between pica and lead exposure is undeniable. It is only the association between pica and mental deficit which has been questioned and it is difficult to understand why this could be the case. If there is an association between mental deficit and lead exposure, and an association between pica and lead exposure, surely there must be an association between mental deficit and pica.
For mental deficit within the normal range, as for mental retardation, the associations between mental ability, pica incidence, and lead exposure, cannot be denied, and there is general acceptance that it is the pica that causes the increase in lead exposure. Quite clearly, these findings are open to the same interpretation as that outlined as applying to the relation of mental retardation to lead exposure, ie, that mental deficit causes pica and pica causes lead exposure. What is lacking in the argument as it applies to mental deficit is evidence that the mental deficit precedes the lead exposure. However, it is implausible to argue that, in mental retardation, the mental defect causes the lead exposure, whereas in mental deficit it is the lead exposure which causes the mental defect. To have one explanation for one and the reverse explanation for the other is quite unreasonable in the absence of evidence that the lead exposure precedes the mental deficit.
The prospective studies: Cord blood lead level and mental deficit
Prospective studies in which lead levels are measured commencing at birth offer the chance of establishing that lead exposure prior to the pica age is related to intelligence as tested later and such a finding is not susceptible to the explanation of reverse causation.
There is no doubt that Needleman and his colleagues had in mind the resolution of the time precedence question when they embarked on the Boston study11 where a large cohort of children was followed from birth. In this study, children were followed from birth with blood samples taken for analysis at birth from cord blood, and then at 6, 12, 18 and 24 months and mental development measured by the Bayley Scales of Infant Development at 12, 18 and 24 months. Three study groups were selected from the babies born at the one hospital according to the lead level in their cord blood: one with levels lower than the 10th percentile; one with levels about the 50th percentile; and one with levels higher than the 90th percentile.
The selection was obviously designed to test the hypothesis that raised cord blood lead levels adversely influenced intelligence as tested later. The study found that infants with raised lead levels in cord blood had poorer scores on the Bayley Scales at 24 months.
The next report of the findings on this cohort of Boston children was made by Bellinger et al (1991)12 after they had been retested at the age of 57 months. It was found that intelligence tested at this time was significantly related to blood lead levels at 24 months, but not related to cord blood lead levels.
The report on this cohort giving the results of retesting at 10 years13 shows, in a table, that cord blood lead levels were not associated with intelligence scores at this age; but this receives no comment in the text.
The claimed association between cord blood lead levels and mental deficit which was originally promoted as the answer to the reverse causation hypothesis has now been discarded by those who were responsible for its promotion. The cord blood hypothesis is not even mentioned by the Boston investigators in their report on testing their cohort at the age of 10 years.
There have been a number of prospective studies following the Boston study and similar in design to it: the Port Pirie study,14,15 the Sydney study,16 the Cincinnati study,3 and the Cleveland study.17 None of these other studies have found any association between cord blood lead levels and intelligence as tested at any age.
The prospective studies: The association found between mental deficit as tested in later years with blood lead levels found at 24 months.
In the Boston study, intelligence as tested at 24 months was not correlated with blood lead levels at this age and the investigators showed little interest in this correlation, the results not being quantified in their first report. Intelligence tested at 57 months was significantly correlated with blood lead level at 24 months (p=0.04), but not with concurrent blood lead levels. Intelligence tested at 10 years was again correlated with blood lead levels at 24 months with even greater statistical significance (p=0.007), but not with levels at any other time.
From this study, a picture has emerged of a special association between blood lead levels at 24 months and intelligence as tested later, with the association becoming more apparent with the lapse of time. This picture finds some support from the other prospective studies; but the picture is not as distinct because these other studies have not reported the associations in terms of blood lead levels at specific ages.
The only finding of significance to emerge from these prospective studies is a negative association between blood lead level at 24 months and intelligence when tested in later years. It may be argued that this finding supports the view that lead exposure is the cause of the mental deficit since the former had precedence in time. It is much more plausible, however, to argue that measures of intelligence used prior to 5 years of age are not valid indicators of adult intelligence.
The point which weighs in favour of the reverse causation hypothesis in these studies is that the intelligence when tested in later years correlates with blood lead levels tested at 24 months but not at any other age. On the conventional hypothesis, with lead exposure as the cause, this finding can only be explained in terms of particular vulnerability of the brain to damage by lead at this age. There is, however, no other evidence for this explanation. On the hypothesis of reverse causation, this finding is precisely what is to be expected from the known peak of blood lead levels due to pica at this age.
Intelligence tests made in infancy do not correlate with intelligence tested in older children or adults, and, it follows that intelligence tests in infancy cannot be used to show that lead exposure causes mental deficit.
We maintain that pica is associated with mental deficit, and this implies that pica in infancy is associated with potential mental deficit, ie, a deficit which can only be assessed at a much later age.
If pica in infants is associated with mental deficit as tested later, it would be expected that this mental deficit should show an association with pica in infants and should also show an association with a surrogate for pica in infants, ie, their blood lead level.
Moreover, it would be expected that this association between mental deficit as measured in older children would correlate most closely with the blood lead level taken at an age when blood lead levels were at their peak under the influence of pica.
This age is almost precisely 24 months.
The argument based on biological plausibility has two aspects, first that high exposure to lead, without encephalopathy, is "known" to cause mental deficit and, secondly, that "lead is a well known neurotoxin".
High exposure to lead, short of encephalopathy, is known to cause mental deficit
Most reviews of the lead question have taken for granted that high exposures to lead are known to cause mental deficit and this implies that low exposures may also do so. The first authors to suggest that lead poisoning without encephalopathy resulted in mental deficit were Byers and Lord (1943).18 Since that time there have been several other studies which have made a similar claim. Without exception, these reports have not controlled for confounding variables such as social class, have no reliable assessment of intelligence status prior to lead exposure and generalise from too few cases. Some of these studies were reviewed by Weiner19 who commented that:
None of the studies provided a definitive answer to the question: Is mental deficiency associated with lead ingestion which is asymptomatic or which produces symptoms less severe than encephalitis?
A rigorous statistical and experimental approach has been conspicuously absent.
The fact is that it has never been shown that lead exposure however high, but short of the production of encephalopathy, produces mental deficit.
Lead is well known to be a neurotoxin
This broad claim is often advanced to promote the biological plausibility of low lead exposures causing mental deficit in children. The human evidence for this argument appears to be based on:
(a) the toxicity to nerves causing peripheral neuritis; and
(b) a belief that lead has a direct destructive effect on the brain neurones which is responsible for the mental retardation following lead encephalopathy.
In large doses lead may cause peripheral neuritis; but few would argue from this that at smaller doses it could cause mental deficit.
In large doses, lead may also cause encephalopathy and sometimes this causes permanent brain damage, manifested in many cases as gross mental retardation. It would be plausible to argue from this that at lower doses lead could cause mental deficit. However, this plausibility is based upon the view that lead causes this damage by a direct effect on cerebral neurones, a view we hold to be incorrect. There is compelling evidence that brain damage resulting from lead encephalopathy is mediated by anoxia; that anoxic brain damage is subject to a threshold is one of the firmest established facts in medical science.
McKhann (1933)20 postulated that all the manifestations of lead encephalopathy including the sequelae of mental retardation, epilepsy, and cerebral palsy were all due to the raised intracranial pressure interfering with the blood supply to the brain. This view implies that brain damage resulting from lead encephalopathy is due to anoxia. McKhann's view was never challenged in medical literature; neither was it supported. Medicine seemed little interested in the pathogenesis of lead encephalopathy at that time. Eventually, however, it seemed to be tacitly agreed that lead encephalopathy was due to direct damage of lead on the neurones. In retrospect, McKhann's view appears much more convincing especially as it has been shown that cerebral anoxia produces the same permanent brain damage as does lead encephalopathy, ie., mental retardation, epilepsy, and cerebral palsy.
For those who would disagree with McKhann, there is other very clear evidence that lead may cause neuronal damage mediated by anoxia. Lead encephalopathy may cause papilloedema and this may be followed by some permanent damage to sight as a result of death of retinal neurones. This neuronal damage is obviously the result of retinal anoxia as it also occurs in raised intracranial pressure due to other causes such as brain tumour or benign intracranial hypertension.21
We take the view that, in lead encephalopathy, brain damage is mediated by anoxia. It is well known that permanent brain damage from anoxia occurs only when the anoxia is sufficiently profound to cause coma. Furthermore, in clinical cases, lead poisoning short of encephalopathy with coma, does not result in permanent brain damage.
On careful examination, the claim that lead has a direct toxic effect on brain neurones cannot be sustained.
The Causation Debate
The hypothesis of reverse causation has been raised by numerous commentators;22-25 yet debate about this has been remarkable for its almost complete absence, although some rather inadequate attempts to clarify the issue have ben made.26,27 This demands an explanation which may simply be given by stating that debate on reverse causation has not been pursued because the debate has been all about whether there is an association or not between lead exposure and mental deficit.
The arguments which have been put forward in support of the view that low level lead exposure causes mental deficit cannot be sustained and the reverse causation hypothesis is a much more plausible explanation of the facts.
1. de la Burde B, Choate MS. Early asymptomatic lead exposure and development at school age. J Pediatr 1972; 81,6: 1088-1091.
2. Needleman HL, Gatsonis CA. Low-level lead exposure and the IQ of children. JAMA 1990; 263, 5:673-678.
3. Dietrich KN, Berger OG, Succop PA, Hammond PB, Bornschein RL. The developmental consequences of low to moderate prenatal lead exposure: Intellectual attainment in the Cincinnati Lead Study cohort following school entry. Neurotoxicol Teratol 1993; 15:37-44.
4. McAlpine C, Singh NN. Pica in institutionalized mentally retarded persons. J Ment Defic Res 1986: 30:171-178.
5. Danford DE, Huber AM. Pica among mentally retarded adults. Am J Ment Def 1982; 87: 141-146.
6. Bayley N. Consistency and variability in the growth of intelligence from birth to eighteen years. J Genet Psychol 1949; 75:165-196.
7. Bornstein MH, Sigman MD. Continuity in mental development from infancy. Child Dev 1986; 57:251-274.
8. Drotar D. Implications of recent advances in neonatal and infant behavioural assessment. J Dev Behav Pediatr 1987; 8:51-53.
9. Moncrieff AA, Koumides OP, Clayton BE, Patrick AG, Renwick AGC, Roberts GE Lead Poisoning in Children. Arch Dis Child 1964; 39:1.
10. Bicknell J, Clayton BE, Delves HT. Lead in mentally retarded children.
J Ment Defic Rec 1968; 12:282-293.
11. Bellinger D, Leviton A, Waternaux C, Needleman H, Rabinowitz M. Longitudinal analyses of prenatal and postnatal lead exposure and early cognitive development. N Eng J Med 1987; 316:1037-1043.
12. Bellinger D, Sloman J, Leviton A, Rabinowitz M, Needleman H, Waternaux C. Low-level lead exposure and children's cognitive function in the preschool years. Paediatrics 1991; 87:219-227.
13. Bellinger D, Stiles K, Needleman H. Low-level lead exposure and academic achievement: A long-term follow-up study. Paediatrics 1992; 90:855-861.
14. McMichael AJ, Baghurst PA, Wigg NR, Vimpani GV, Robertson EF, Roberts RJ et al. Port Pirie Cohort Study: Environmental exposure to lead and children's abilities at the age of four years. N Eng J Med 1988; 319:468-475.
15. Baghurst P, McMichael AJ, Wigg NR, Vimpani GV, Robertson EF, Roberts RJ et al. The Port Pirie Cohort Study: Environmental exposure to lead and children's intelligence at the age of seven years. N Eng J Med 1992; 327:1279-1284.
16. Cooney G, Bell A, Stavrou C. Low level exposure to lead and neurobehavioural development: the Sydney study at seven years. Paper presented at the Heavy Metal Conference, Edinburgh 1992.
17. Ernhart CB, Morrow-Tlucak M, Wolf AW, Super D, Drotar D. Low level lead exposure in the prenatal and early preschool periods: Intelligence prior to school entry. Neurotoxicol Teratol 1989; 11:161-170.
18. Byers RK, Lord EE. Late effects of lead poisoning on mental development. Am J Dis Child 1943; 66:471-494.
19. Weiner G. Varying Psychological Sequelae of Lead Ingestion in Children.
Public Health Rep 1970; 85:19-24.
20. McKhann CF, Vogt EC. Lead poisoning in children. JAMA 1933; 101:1131-1135.
21. Grant DN. Benign intracranial hypertension. Arch Dis Child 1971; 46:651-655.
22. Clayton BE. Lead: the relation of environment and experimental work.
Br Med Bull 1975; 31:230-240.
23. Rutter M. Raised lead levels and impaired cognitive/behavioural functioning: a review of the evidence. Devel Med Child Neurol (Suppl) 1980; 22:1-26.
24. Smith M Delves T, Lansdown R, Clayton B, Graham P. The effects of lead exposure on urban children: The institute of Child Health/Southhampton Study, Dev Med Child Neurol 1983; 25:1-54.
25. The Medical Research Council (1984). The neurological effects of lead in children, A review of recent research 1979-83.
26. Christophers AJ. Environmental lead question. A point of view from Victoria. Med J Aust 1980; 2:300-304.
27. Cole JF. Letter to the Editor. New Eng J Med 1979; 30: 161-162.11